notes/http-and-webdav/webdav-usage.html

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<style type="text/css"> /* <![CDATA[ */
  @import "branding/css/tigris.css";
  @import "branding/css/inst.css";
  /* ]]> */</style>
<link rel="stylesheet" type="text/css" media="print"
  href="branding/css/print.css"/>
<script type="text/javascript" src="branding/scripts/tigris.js"></script>
<title>Use of WebDAV in Subversion</title>
</head>

<body>
<div class="app">

    <h2>Use of WebDAV in Subversion</h2>

    <p>
      This document details how WebDAV is used within the
      <a href="http://subversion.apache.org/">Subversion
      product</a>. Specifically, how the client side interfaces with
      <a href="http://www.webdav.org/neon/">Neon</a> to generate
      WebDAV requests over the wire, and what the
      server must do to map incoming WebDAV requests into operations
      against the Subversion repository. Note that the server side is
      implemented as an
      <a href="http://httpd.apache.org/">Apache 2.0</a> module,
      operating as a back-end for its mod_dav functionality.
    </p>
    <p>
      This document heavily refers to the
      <a href="http://svn.apache.org/repos/asf/subversion/trunk/notes/subversion-design.html">Subversion
	design document</a> and the
      <a href="http://www.webdav.org/deltav/">latest Delta-V protocol
	draft</a>. Details of those documents will <em>not</em> be
      replicated here.
    </p>

      <table width="70%" border="0" cellspacing="0" cellpadding="10"
      style="background-color: #ff0000; color: white;
      margin-left: auto; margin-right: auto">
	<tr>
	  <td>
	      <strong>NOTE:</strong> Subversion uses DeltaV for its
	      communication, but the Subversion client is
	      <strong>not</strong> a general-purpose DeltaV client. In
	      fact, it expects some custom features from the
	      server. Further, the Subversion server is
	      <strong>not</strong> a general-purpose DeltaV server. It
	      implements a strict <strong>subset</strong> of the
	      DeltaV specification. A WebDAV or DeltaV client may very
	      well be able to interoperate with it, but only if that
	      client operates within the narrow confines of those
	      features the server has implemented.
	  </td>
	</tr>
	<tr>
	  <td>
	      Version 2.0 of Subversion will address WebDAV
	      interoperability (Class 1, Class 2, and DeltaV
	      features). Each of the custom features expected by the
	      client actually has an alternate mechanism available in
	      DeltaV, but in a much less efficient form.
	  </td>
	</tr>
	<tr>
	  <td>
	      It is expected that Version 1.0 will support read-only,
	      Class 1 WebDAV clients. Any "low-hanging fruit" to
	      increase DeltaV interoperability will be considered.
	  </td>
	</tr>
      </table>

    <p></p>

    <h3>Basic Concepts</h3>
    <p>
      Subversion uses a tree-based format to describe a change set
      against the repository. This tree is constructed on the client
      side (by "walking" over the working copy) to describe the
      change. The tree is marshalled to the server as a linear
      sequence of changes to be applied to the repository. The
      repository can accept changes in a random-access manner, so the
      mapping from a tree to a set of changes works very well for the
      repository.
    </p>
    <p>
      Subversion provides properties on files, directories, and even
      the abstract concept of a revision. Each of the operations
      involving properties are mapped directly to WebDAV properties,
      which are manipulated with the <span
      class="method">PROPFIND</span> and <span
      class="method">PROPPATCH</span> HTTP methods.  Revisions are
      modeled as DeltaV <span class="term">baselines</span>, so
      revision properties are available through a <span
      class="method">PROPFIND</span> on the baseline.
    </p>
    <p>
      The Subversion server can efficiently compute deltas between two
      revisions (these deltas are complete <span class="term">tree
      deltas</span>, not simple text deltas). DeltaV does not have a
      direct analogue for the tree delta concept. A client could
      discover changes by issuing a sequence of <span
      class="method">PROPFIND</span> requests on the various WebDAV
      resources, but this would be a time-consuming operation,
      involving many requests. Instead, Subversion marshals this
      concept as a custom WebDAV <span
      class="term">report</span>. Using this report, the client learns
      which items in the working copy are out of date and can issue
      <span class="method">GET</span> and <span
      class="method">PROPFIND</span> methods to fetch the new data.
    </p>
    <p>
      Tags and branches are simple copies in Subversion, which are
      handled with the WebDAV <span class="method">COPY</span>.
    </p>
    <blockquote class="comment">
      <p>need to talk about copies somewhere. need to discuss how copy
      history is retained (svn does it automatically, but interop with
      other servers may require us to set a custom property on those
      servers.</p>
    </blockquote>

    <h3>DeltaV Concepts Used by Subversion</h3>
    <p>
      Subversion uses many of the DeltaV concepts, as listed
      below. Note that many of these concepts are not fully
      implemented by Subversion; we have implemented enough to meet
      our needs, but little more.
    </p>
    <dl>
      <dt>Baseline</dt>
      <dd>
	<span class="comment">further info to come...</span><p></p>
      </dd>

      <dt>Activity</dt>
      <dd>
	<span class="comment">further info to come...</span><p></p>
      </dd>

      <dt>Version Resource</dt>
      <dd>
	<span class="comment">further info to come...</span><p></p>
      </dd>

      <dt>Version-Controlled Configuration</dt>
      <dd>
	<span class="comment">further info to come...</span><p></p>
      </dd>

      <dt>Baseline Collection</dt>
      <dd>
	<span class="comment">further info to come...</span><p></p>
      </dd>

      <dt>Version-Controlled Resource</dt>
      <dd>
	<span class="comment">further info to come...</span><p></p>
      </dd>

      <dt>Working Resource (Feature)</dt>
      <dd>
	<span class="comment">further info to come...</span><p></p>
      </dd>

      <dt>Merge Feature</dt>
      <dd>
	<span class="comment">further info to come...</span><p></p>
      </dd>

      <dt>Label Feature</dt>
      <dd>
	<span class="comment">further info to come...</span><p></p>
      </dd>

      <dt>Version-Controlled-Collection Feature</dt>
      <dd>
	<span class="comment">further info to come...</span><p></p>
      </dd>

    </dl>

    <h3>Subversion Projects as URLs</h3>
    <p>
      The very first concept to define is how a project is exposed to
      the client. Subversion will expose all projects as URLs on a
      server. The files and subdirectories under this project will be
      exposed through the URL namespace.
    </p>
    <p>
      For example, let us assume that we have a project named
      "example". And let us say that this project will be exposed at
      the URL:
      <span class="url">http://subversion.tigris.org/repos/example/</span>.
    </p>
    <p>
      This mapping is set up through a set of configuration
      parameters for the Apache HTTP Server (which is hosting the
      Subversion code and the particular project in question). The
      configuration could look like:
    </p>
    <blockquote>
<pre>&lt;Location /repos/example&gt;
    DAV svn
    SVNPath /home/svn-projects/example
&lt;/Location&gt;</pre>
    </blockquote>
    <p>
      Files and directories within the project will be directly mapped
      into the URL namespace. For example, if the project contains a
      file "file.c" in a subdirectory "sub", then the URL for that
      file will be
      <span class="url">http://subversion.tigris.org/repos/example/sub/file.c</span>.
    </p>

    <h3>Initial Checkout</h3>
    <p>
      When the user performs the initial checkout of a Subversion
      project, the client will issue a series of <span
      class="method">PROPFIND</span> and <span
      class="method">GET</span> requests. These requests will traverse
      the repository, pick up some necessary metadata, and then fetch
      the latest revision.
    </p>
    <p class="comment">
      describe the OPTIONS request for fetching the activity
      collection set. describe the sequence of PROPFINDs to reach the
      baseline collection.
    </p>
    <p class="comment">
      <i>(moved here from below; need to rewrite)</i><br/>
      When the initial checkout was performed, Subversion fetched
      the <span class="prop">DAV:activity-collection-set</span> value
      and stored it as a property on each directory in the working copy.
property for each
      collection. This property lists all of the locations on the
      server where an activity may be created. The first of these
      locations will be stored on the client for use during the commit 
      process.
    </p>
    <p class="comment">
      Should probably describe the metadata we fetch, and how a
      checkout of "not the latest" (e.g. by date or revision) will
      work.
    </p>

    <h3>Committing a Change</h3>
    <p>
      Subversion commits are modeled using the "activity" concept from
      DeltaV. An activity can be viewed as a transaction for a set of
      resources.
    </p>

    <h3>Creating the activity</h3>
    <p>
      At commit time, the Subversion client will retrieve the stored
      <span class="prop">DAV:activity-collection-set</span> value to
      know where it should create the activity. Next,
      the client will generate a UUID (a unique value) to use for the
      activity's location. Finally, the client will issue a
      <span class="method">MKACTIVITY</span> method request, where the Request-URL is
      composed from the activity location and the UUID. This request
      will construct an activity to hold all of the changes for the
      commit.
    </p>
    <p>Abbreviated summary:</p>
    <dl>
      <dt>At checkout time:</dt>
      <dd>
	<div>Request: <span class="method">OPTIONS</span> for
	<span class="prop">DAV:activity-collection-set</span></div>
        <div>Response: <span
        class="url">http://www.example.com/repos/foo/$svn/act/</span></div>
        </dd>
      <dt>At commit time:</dt>
      <dd>
	<div>Request: <span class="method">MKACTIVITY</span>
	<span class="url">http://www.example.com/repos/foo/$svn/act/01234567-89ab-cdef-0123-456789abcdef</span></div>
	<div>Response: 201 (Created)</div>
      </dd>
    </dl>
    <p>
      The <span class="method">CHECKOUT</span> method can specify an
      activity to use upon checkout. This feature is used to associate
      all items with the newly-created activity.
    </p>

    <h3>Storing the commit message</h3>
    <p class="comment">
      talk about checking out the baseline and applying a PROPPATCH to
      the working baseline.
    </p>

    <h3>Mapping changes to WebDAV</h3>
    <p>
      A change set in Subversion is specified with a "tree delta" (see
      the SVN design for more details on the changes that can be
      placed into a tree delta). The tree delta will be unravelled
      into a set of requests. These requests will be one of the
      following forms:
    </p>
    <dl>
      <dt>Delete file or directory</dt>
      <dd>
	These changes are mapped onto a <span
	class="method">DELETE</span> operation. The version resource
	of the target's parent collection is checked out using the
	<span class="method">CHECKOUT</span> method (into the current
	activity). The target (name) is then deleted from the
	resulting working collection using the <span
	class="method">DELETE</span> method.

	<p></p>
      </dd>

      <dt>Add file</dt>
      <dd>
	This is modeled by performing a <span
	class="method">CHECKOUT</span> of the version resource of the
	target's parent collection. The new file is created within the
	resulting working collection using a <span
	class="method">PUT</span> request. Properties are applied
	using <span class="method">PROPPATCH</span>.

	<p></p>
      </dd>

      <dt>Add directory</dt>
      <dd>
	This is modeled by performing a <span
	class="method">CHECKOUT</span> on the version resource of the
	target's parent collection. The new directory is created
	within the resulting working collection with a <span
	class="method">MKCOL</span> request. Properties are applied
	using <span class="method">PROPPATCH</span>.

	<p></p>
      </dd>

      <dt>Add file or directory, with previous ancestory (a copy)</dt>
      <dd>
	<span class="comment">need to fix this section</span>

	<p>
	  A tree delta can specify that a file/directory originates as
	  a copy of another file/dir. This copy may be further
	  modified by additional elements the tree delta.
	</p>
	<p>
	  This change will be modeled by performing a
	  <code>CHECKOUT</code> on the version resource of the parent
	  collection which will contain the new resource. The
	  <code>VERSION-CONTROL</code> method will create a new
	  version-controlled resource (VCR) within the working
	  collection, with the VCR's <code>DAV:checked-in</code>
	  property referring to the ancestor's version resource.
	</p>
	<blockquote>
	  <p><strong>Note:</strong> it appears that we will use
	  <code>COPY</code> to copy the appropriate resource into the
	  working collection. This will create a new version history
	  which is then placed into the working collection. The
	  version history will use the <code>DAV:precursor-set</code>
	  property to specify the version resource of the ancestor.</p>

	  <p>
	    Because a version resource does not specify the revision,
	    it will not be possible to <code>COPY</code> a version
	    resource into the working collection -- it will not tell
	    us what revision was copied. Instead, we will most likely
	    copy a version resource out of the appropriate
	    baseline. This implies the client must be able to map from 
	    a URL/revision pair to a baselined version resource URL.
	  </p>
	  <p>
	    The second issue is whether/how we set the
	    <code>DAV:precursor-set</code> property of the version
	    history. Or, more precisely, how we synthesize the value
	    from information stored in the repository. This is still
	    under investigation.
	  </p>
	</blockquote>
      </dd>

      <dt>Replace file/dir by another file/dir</dt>
      <dd>
	This change does not have a WebDAV modeling because tree
	deltas model it as two, sequential operations: a
	<em>delete</em>, followed by an <em>add</em>.

	<p></p>
      </dd>

      <dt>Moving a file or directory</dt>
      <dd>
	This change does not have a WebDAV modeling because tree
	deltas model it as two, distinct operations: a
	<em>delete</em>, and an <em>add</em> with previous ancestry.

	<p></p>
      </dd>

      <dt>Replace file</dt>
      <dd>
	This is modeled with a <span class="method">CHECKOUT</span> on
	the target's version resource, followed by a <span
	class="method">PUT</span> to the resulting working resource.

	<p></p>
      </dd>

      <dt>Replace directory</dt>
      <dd>
	In Subversion terms, "replace directory" means that additions, 
	deletions, and other changes will occur <em>within</em> the
	directory. Each of these changes are modeled individually, and 
	the change to the directory is performed
	implicitly. Therefore, this "change" has no particular mapping 
	into WebDAV.

	<p></p>
      </dd>

      <dt>Property delta</dt>
      <dd>
	A property delta (against a file or directory) maps directly
	to a <span class="method">PROPPATCH</span> in WebDAV
	terms. The target's version resource will be checked out using
	<span class="method">CHECKOUT</span> and the <span
	class="method">PROPPATCH</span> will be applied to the
	resulting working resource.
      </dd>
    </dl>

    <h3>Final Commit</h3>
    <p>
      The final action of the commit process is to issue a <span
      class="method">MERGE</span> request to the Subversion server,
      specifying that the activity (created earlier) be checked in and
      the corresponding version-controlled resources be updated to
      refer to the new version resources.
    </p>
    <p class="comment">
      the comment below is not quite right. talk about the working
      baseline, and how that is used to create a new baseline (with
      the commit message on it)
    </p>
    <p>
      The version-controlled resources are also baseline-controlled,
      which means that updates to them will automatically create a new 
      baseline. In essence, the commit will create a new baseline
      corresponding to the new Subversion revision.
    </p>

    <h3>Example</h3>
    <p class="comment">
      <strong>Warning:</strong> this section has not been updated to
      reflect some recent changes to the SVN-to-DAV mapping. Consider
      it out of date until this warning is removed.
    </p>
    <p>
      Consider the following set of operations and its corresponding
      tree delta (taken from the SVN design document):
    </p>
    <ol>
      <li>rename <code>/dir1/dir2</code> to <code>/dir1/dir4</code>,</li>
      <li>rename <code>/dir1/dir3</code> to <code>/dir1/dir2</code>, and</li>
      <li>move <code>file3</code> from <var>/dir1/dir4</var> to <var>/dir1/dir2</var>.</li>
    </ol>
<pre>&lt;tree-delta&gt;
  &lt;replace name='dir1'&gt;
    &lt;directory&gt;
      &lt;tree-delta&gt;
        &lt;replace name='dir2'&gt;
          &lt;directory ancestor='/dir1/dir3'&gt;         (1)
            &lt;tree-delta&gt;
              &lt;new name='file3'&gt;                    (2)
                &lt;file ancestor='/dir1/dir2/file3'/&gt;
              &lt;/new&gt;
            &lt;/tree-delta&gt;
          &lt;/directory&gt;
        &lt;/replace&gt;
        &lt;delete name='dir3'/&gt;                       (3)
        &lt;new name='dir4'&gt;                           (4)
          &lt;directory ancestor='/dir1/dir2'&gt;
            &lt;tree-delta&gt;
              &lt;delete name='file3'/&gt;                (5)
            &lt;/tree-delta&gt;
          &lt;/directory&gt;
        &lt;/new&gt;
      &lt;/tree-delta&gt;
    &lt;/directory&gt;
  &lt;/replace&gt;
&lt;/tree-delta&gt;
</pre>

    <p>
      Walking through this delta, we map out the WebDAV requests
      listed below. The numbers in the above delta roughly correspond
      to the numbered entries below. The correspondence is not exact
      because a specific, resulting behavior is typically based on a
      combination of a few elements in the delta.
    </p>
    <ol>
      <li>
	The <code>&lt;directory ancestor="/dir1/dir3"&gt;</code>
	specifies that we are overwriting <code>/dir1/dir2</code> with 
	<code>/dir1/dir3</code>.
	<p>
	  <code>CHECKOUT&nbsp;/dir1/dir2/</code><br/>
	  <i>(returns a working resource URL for the directory)</i>
	</p>
	<p>
	  <code>COPY&nbsp;/dir1/dir3/</code><br/>
	  <code>Destination:&nbsp;http://www.example.com/$svn/wrk/.../</code><br/>
	  <code>Overwrite:&nbsp;T</code>
	</p>
      </li>
      <li>
	<code>/dir1/dir2/file3</code> is new (since we just overwrote
	the original <code>dir2</code> directory), and originates from
	<code>/dir1/dir2/file3</code>. Thus, we simply
	<code>COPY</code> the file into the target directory's working 
	resource:
	<p>
	  <code>COPY&nbsp;/dir1/dir2/file3</code><br/>
	  <code>Destination:&nbsp;http://www.example.com/$svn/wrk/.../file3</code>
	</p>
      </li>
      <li>
        <p>
          <code>CHECKOUT&nbsp;/dir1/dir3/</code><br/>
          <i>(returns a working resource URL for the directory)</i>
        </p>
	<p>
	  <code>DELETE&nbsp;/$svn/wrk/.../</code>
	</p>
      </li>
      <li>
	We are going to  creating a new subdirectory (<code>dir4</code>) in the
	<code>/dir1</code> directory. Since we don't have
	<code>/dir1</code> checked out yet, we do so:
	<p>
	  <code>CHECKOUT&nbsp;/dir1/</code><br/>
	  <i>(returns a working resource URL for the directory)</i>
	</p>
	<p>
	  And now we copy the right directory into the new working
	  resource:
	</p>
	<p>
	  <code>COPY&nbsp;/dir1/dir2/</code><br/>
	  <code>Destination:&nbsp;http://www.example.com/$svn/wrk/.../dir4/</code>
	</p>
      </li>
      <li>
	The <code>COPY</code> created a complete set of working
	resources on the server, so we simply delete the part that we
	don't want:
	<p>
	  <code>DELETE:&nbsp;/$svn/wrk/.../dir4/file3</code>
	</p>
      </li>
    </ol>

    <h3>URL Layout</h3>
    <p>
      The Subversion server exposes repositories at user-defined
      URLs. For example, the "foo" repository might be located at
      <span
      class="url">http://www.example.com/repos/foo/</span>. However,
      the server also requires a number of other resources to be
      exposed for proper operation. These additional resources will be
      associated with each repository in a location under the main
      repository URL. By default, this location is "<span
      class="url">$svn</span>". It may be changed by using the
      <code>SVNSpecialURI</code> directive:
    </p>
    <blockquote>
<pre>&lt;Location /repos/foo&gt;
    DAV svn
    SVNPath /home/svn-projects/foo
    SVNSpecialURI .special
&lt;/Location&gt;</pre>
    </blockquote>
    <p>
      Underneath the location specified by <code>SVNSpecialURI</code>,
      we will expose several collections. Assuming we use the default
      of "<span class="url">$svn</span>", the collections are:
    </p>
    <dl>
      <dt><span class="url">$svn/act/</span></dt>
      <dd>
	This area is where activity resources are created. The client
	will pick a unique name within this collection and issue a
	<span class="method">MKACTIVITY</span> for that URL. The client will then use 
	the activity in further interactions.
	<p>
	  No methods are allowed on the <span class="url">$svn/act/</span>
	  resource.
	</p>
	<blockquote>
	  <p>Note: actually, we may want to allow a <code>PROPFIND</code> 
	  with a <code>Depth:&nbsp;1</code> header to allow clients to 
	  enumerate the current activities.</p>
	</blockquote>
	<p>
	  Only a subset of methods are allowed on the activities
	  within the collection. They are: <span class="method">PROPFIND</span>,
	  <span class="method">MERGE</span> (commit the activity), and
	  <span class="method">DELETE</span> (abort the activity).
	</p>
	<p>
	  Per the Delta-V specification, all activity resources will
	  have a <span class="prop">DAV:resourcetype</span> of
	  <span class="prop">DAV:activity</span>.
	</p>
      </dd>
      
      <dt><span class="url">$svn/his/</span></dt>
      <dd>
	<span class="comment">do something with this section; we
	actually don't use version history resources. in the future,
	they might be modeled like this</span>

	<p>
	  This collection contains the version history resources for
	  files and directories in a project. Its internal layout is
	  completely server-defined. Clients will receive URLs into
	  this collection (or a subcollection) from various responses.
	</p>
	<p>
	  No methods are allowed on the <span class="url">$svn/his/</span>
	  resource.
	</p>
	<p>
	  Internally, the URL namespace is laid out with URLs of the
	  following form:
	</p>
	<blockquote class="url">
	  <p>$svn/his/<var>node-id</var></p>
	</blockquote>
	<p>
	  The <var>node-id</var> is an internal value
	  that Subversion uses to reference individual files and
	  directories. This <var>node-id</var> is a single integer
	  defined by the Subversion repository. Note that this is an 
	  undotted node id, which is the base for the entire history
	  of a given node in the repository.
	</p>
	<p>
	  The <span class="prop">DAV:resourcetype</span> of the <var>node-id</var>
	  collection is <span class="prop">DAV:version-history</span>.
	</p>
	<blockquote class="comment">
	  <p><strong>Note:</strong> the above information is probably not 
	  quite correct. The issue of linking one version history to
	  another is still open. Further, I think that node 73 and
	  node 73.4.1 are each version histories (where the latter is
	  linked to the former). 73.x and 73.4.1.x are the versions
	  within the version history.</p>
	</blockquote>
      </dd>

      <dt><span class="url">$svn/ver/</span></dt>
      <dd>
	This collection contains the version resources for the
	project.

	<p>
	  No methods are allowed on the <span class="url">$svn/ver/</span>
	  resource.
	</p>
	<p>
	  The layout of this collection is internal to the server. For
	  reference purposes here (and to describe the
	  implementation), it is laid out as:
	</p>
	<blockquote class="url">
	  <p>$svn/ver/<var>node-id</var>/<var>path</var></p>
	</blockquote>
	<p>
	  Only read-only methods are allowed against these resources
	  (e.g. <span class="method">GET</span>, <span class="method">PROPFIND</span>,
	  <span class="method">REPORT</span>); all other methods are illegal.
	</p>
	<p>
	  The <span class="prop">DAV:resourcetype</span> of a version resource is
	  simply the value of the resource at checkin time
	  (e.g. <code>&lt;D:resourcetype/&gt;</code> or
	  <code>&lt;D:resourcetype&gt;&lt;D:collection/&gt;&lt;/D:resourcetype&gt;</code>).
	</p>
      </dd>

      <dt><span class="url">$svn/wrk/</span></dt>
      <dd>
	This collection contains working resources for the resources
	that have been checked out with the <span class="method">CHECKOUT</span>
	method. The form and construction of this collection is
	server-defined, but is also well-defined so that clients may
	interact properly with collection versions that have been
	checked out.
	<p>
	  No methods are allowed on the <span class="method">$svn/wrk/</span>
	  resource.
	</p>
	<p>
	  For reference purposes, the working resource URLs are
	  constructed as:
	</p>
	<blockquote class="url">
	  <p>$svn/wrk/<var>activity</var>/<var>path</var></p>
	</blockquote>
	<p>
	  Any method is allowed on the working resources, but no
	  methods are allowed on any of its parents.
	</p>
	<p>
	  The <span class="prop">DAV:resourcetype</span> of the working resources
	  follows normal resource typing:
	  <code>&lt;D:resourcetype/&gt;</code> for regular working
	  resources, and
	  <code>&lt;D:resourcetype&gt;&lt;D:collection/&gt;&lt;/D:resourcetype&gt;</code>
	  for working collections.
	</p>
      </dd>

      <dt><span class="url">$svn/vcc/</span></dt>
      <dd>
	<span class="comment">This section is not yet complete.</span>

	<p>
	  version-controlled configuration...
	</p>

	<p>
	  <code>$svn/vcc/root</code> as a singleton.
	</p>
      </dd>

      <dt><code>$svn/bln/</code></dt>
      <dd>
	<span class="comment">This section is not yet complete.</span>

	<p>
	  baselines...
	</p>

	<p>
	  <code>$svn/bln/<var>rev</var>/</code>
	</p>
      </dd>

      <dt><code>$svn/wbl/</code></dt>
      <dd>
	<span class="comment">This section is not yet complete.</span>

	<p>
	  working baseline...
	</p>
      </dd>

      <dt><code>$svn/bc/</code></dt>
      <dd>
	<span class="comment">This section is not yet complete.</span>

	<p>
	  baseline collection...
	</p>
      </dd>
    </dl>

    <h3>Property Management (and History/Log Reporting)</h3>
    <p class="comment">
      this section needs to be reworked. the properties occur on the
      FS revisions (and exposed via baselines).
    </p>
    <p>
      As mentioned before, Subversion properties map onto WebDAV
      properties. For history/log reporting, the following WebDAV
      properties will be applied to each baseline (a Subversion
      revision) and to each version resource created by the
      revision. Since these resources are all version resources, the
      properties below are read-only.
    </p>
    <dl>
      <dt><code>DAV:comment</code></dt>
      <dd>
	This is the standard (dead) property for specifying a checkin
	comment.
	<p></p>
      </dd>
      <dt><code>DAV:creator-displayname</code></dt>
      <dd>
	This is a (dead) property that is generated from Subversion's
	concept of the "user" who made a particular change.
	<p></p>
      </dd>
      <dt><code>DAV:creationdate</code></dt>
      <dd>
	This is a read-only live property created by the server at
	commit time.
      </dd>
    </dl>
    <p>
      The history for a specified file will be generated using the
      <code>REPORT</code> method and a
      <code>DAV:property-report</code> report. A typical history will
      fetch the three properties mentioned above for each version of
      the file/directory.
    </p>
    <p>
      Based on the client design, it may be important to specify other 
      read-only live properties for information about versions. For
      example, how many lines were added/removed in a particular
      checkin for a file? Creating these live properties will be quite 
      straight-forward, and driven by the client design over time.
    </p>
    <blockquote>
      <p>Note: if we do this, however, then we'd end up tying the client
      to the server. Of course, if the client were run against another 
      DeltaV server which didn't report these properties, then we'd
      simply not display them in the UI. (e.g. graceful degradation of 
      functionality)</p>
    </blockquote>

    <h3>Fetching Status and Updates</h3>
    <p>
      After the initial checkout, the client can request a status
      report (what has been changed on the client, pending a commit;
      what has been changed on the server, pending an update). The
      update process is similar, except that we also fetch the changes 
      from the server.
    </p>
    <p>
      The local changes can be handled entirely on the client
      side. The Working Copy library can easily handle the detection
      and reporting of these changes. We're concerned with efficiently 
      detecting what has changed on the server.
    </p>
    <p>
      While it would be possible to traverse the repository, fetching
      the current state, and comparing that to the client state, it
      would not be efficient. The Subversion design enables the server
      to easily compute what has changed (relative to the client), if
      it is given a description of the client state.
    </p>
    <p>
      The core of the <em>status</em> and <em>update</em> commands is
      based on a custom Subversion-specific WebDAV report. This custom
      report will transmit the state of the working copy to the
      server, and the server response will specify which resources
      will need to be updated (fetched).
    </p>
    <p>
      The request is a standard <code>REPORT</code> request, with a
      custom XML body. The body will use the standard Subversion
      technique of reporting a top-level revision number, and then
      only reporting children that have different revisions. The
      result of the report will use the same technique of reporting
      only the resources where a change is found. If a change is
      found, the server will provide a URL to the version resource to
      fetch for the changed resource. The server will also report the
      current revision number.
    </p>
    <blockquote class="comment">
      <p>The XML DTDs for the request and response are TBD.</p>
    </blockquote>
    <blockquote>
      <p><small>
	The custom report will tie the client to only those servers
	which support the report, but a future version of the software
	will contain a fallback codepath, a graceful degradation, to
	support other DeltaV servers.
      </small></p>
    </blockquote>
    <p>
      When an updated is performed, the client will fetch each of the
      URLs (using <code>GET</code> requests) provided in the server
      response.
    </p>
    <p>
      <code>GET</code> (and <code>PUT</code>) operations will transfer 
      content in a "diff" format when possible. The mechanics of this
      will follow the Internet Draft, titled
      <a href="http://www.ietf.org/internet-drafts/draft-mogul-http-delta-10.txt">Delta Encoding in HTTP</a>.
    </p>

    <h3>Entity Tags (etags)</h3>
    <p>
      Etags are required to be unique across all versions of a
      resource. Luckily, this
      is very easy for a version control system. Each etag will be
      simply be the repository's <em>node-id</em> for the resource.
    </p>
    <p>
      Etags are used to generate diffs, following the guidelines in
      the aforementioned draft:
      <a href="http://www.ietf.org/internet-drafts/draft-mogul-http-delta-10.txt">Delta Encoding in HTTP</a>.
      The problem then becomes how to get the etag for each file
      stored on the client (we don't need etags for directories since
      we never fetch them). During a <em>checkout</em> or
      <em>update</em> process, this is easy: the etag is provided in
      the HTTP response headers for each file retrieved.
    </p>
    <p>
      The other part of the problem is getting the etag after a
      <em>commit</em> has occurred. The <code>MERGE</code> response
      provides a way to request properties from the version resources
      which are created as part of the checkin of the activity. The
      etag (and other properties) can be fetched using that mechanism.
    </p>

    <h3>Tags and Branches</h3>
    <p>
      Tags and branches within Subversion are performed by copying
      from one area to another. For example:
    </p>
    <blockquote>
      <pre>
[.../src/my-project]$ svn cp trunk tags/1.0.3-rc4
[.../src/my-project]$ svn commit
</pre>
    </blockquote>
    <p>
      In the above example, <code>tags/1.0.3-rc4</code> should now be
      considered readonly and will always reflect the status of
      <code>trunk</code>.
    </p>
    <p>
      These copies are handled just like a regular commit. An activity
      is created with <span class="method">MKACTIVITY</span>, a
      working resource is created via <span
      class="method">CHECKOUT</span> (for the target directory;
      <code>tags/</code> in our example above), and then a <span
      class="method">COPY</span> is performed. The activity is then
      merged back into the repository with a <span
      class="method">MERGE</span> request.
    </p>

    <h3>Server Requirements</h3>
    <p class="comment">
      <strong>Warning:</strong> this section is out of date. The
      DeltaV draft has gone through a number of revisions, and our use
      of DeltaV has changed some.
    </p>

    <h3>DAV Methods</h3>
    <p>
      The server will need to implement the following WebDAV methods
      for proper operation:
    </p>
    <ul>
      <li><span class="method">OPTIONS</span></li>
      <li><span class="method">GET</span></li>
      <li><span class="method">DELETE</span></li>
      <li><span class="method">COPY</span></li>
      <li><span class="method">PROPPATCH</span></li>
      <li><span class="method">PROPFIND</span></li>

      <li><span class="method">MKACTIVITY</span></li>
      <li><span class="method">CHECKOUT</span></li>
      <li><span class="method">MERGE</span></li>
      <li><span class="method">REPORT</span></li>
    </ul>
    <p>
      The following methods are not required by Subversion at this
      time:
    </p>
    <ul>
      <li><span class="method">CHECKIN</span></li>
      <li><span class="method">UNCHECKOUT</span></li>
      <li><span class="method">UPDATE</span></li>
      <li><span class="method">LABEL</span></li>
      <li><span class="method">VERSION-CONTROL</span></li>
      <li><span class="method">BASELINE-CONTROL</span></li>
      <li><span class="method">MKWORKSPACE</span></li>
    </ul>

    <h3>DAV Properties</h3>
    <p>
      The following DeltaV properties will be implemented:
    </p>
    <ul>
      <!-- resource properties -->
      <li><span class="prop">DAV:comment</span></li>
      <li><span class="prop">DAV:creator-displayname</span></li>
      <li><span class="prop">DAV:supported-method-set</span></li>
      <li><span class="prop">DAV:supported-live-property-set</span></li>
      <li><span class="prop">DAV:supported-report-set</span></li>
      <li><span class="prop">DAV:version-controlled-configuration</span></li>

      <!-- version-controlled resource properties -->
      <li><span class="prop">DAV:checked-in</span></li>
      <li>
	<span class="prop">DAV:auto-version</span> is a readonly,
	empty element (auto versioning not supported).
      </li>

      <!-- checked-out resource properties -->
      <li><span class="prop">DAV:checked-out</span></li>
      <li>
	<span class="prop">DAV:predecessor-set</span>
        <div><i>
	  Note: the Subversion design document is not clear on the
	  mechanics of how multiple predecessors are merged to create
	  a single, new revision. When this clarifies, then
	  <code>DAV:predecessor-set</code> may end up containing more
	  than zero or one predecessor URLs
	</i></div>
      </li>

      <!-- version resource properties -->
      <!-- predecessor-set -->
      <li>
	<span class="prop">DAV:version-name</span> is simply the
	(global) revision number.
      </li>
      <li><span class="prop">DAV:checkout-fork</span></li>
      <li><span class="prop">DAV:checkin-fork</span></li>

      <!-- checked-out (working resource) properties -->
      <!-- checkout-fork, checkin-fork -->
      <li><span class="prop">DAV:auto-update</span></li>
      <li>
	<span class="prop">DAV:subbaseline-set</span> is a readonly,
	empty property (sub-baselines not supported).
      </li>
      <li>
	<span class="prop">DAV:unreserved</span> is set to
	<code>F</code>.
      </li>

      <!-- vcc properties -->
      <li><span class="prop">DAV:baseline-controlled-collection</span></li>

      <!-- baseline properties -->
      <!-- subbaseline-set -->
      <li><span class="prop">DAV:baseline-collection</span></li>

      <!-- activity properties -->
      <li>
	<span class="prop">DAV:subactivity-set</span> is a readonly,
	empty property (sub-activities not supported).
      </li>

      <!-- version-controlled collection properties -->
      <li>
	<span class="prop">DAV:eclipsed-set</span> is always empty
	(internal members can never be eclipsed).
      </li>
    </ul>

    <p>
      Contrary to the DeltaV specification, the following required
      properties will not be implemented:
    </p>
    <ul>
      <li>
	<span class="prop">DAV:successor-set</span> - this may be an
	expensive operation to synthesize this value.
      </li>
      <li>
	<span class="prop">DAV:checkout-set</span> - we do not record
	what has actually been checked out, but use the working
	resource URL to provide the necessary information; thus we
	have no record of the data to populate this property.
      </li>
      <li>
	<span class="prop">DAV:merge-set</span> - our <span
	class="method">MERGE</span> method is solely to support a
	commit. It doesn't really support the arbitrary merging
	defined within the spec.
      </li>
      <li>
	<span class="prop">DAV:auto-merge-set</span> - same as for
	<span class="prop">DAV:merge-set</span>.
      </li>
      <li>
	<span class="prop">DAV:activity-version-set</span> -
	activities are only used for working resources, so versions
	cannot be part of an activity.
	<div><i>maybe this should be defined as the empty set?</i></div>
      </li>
      <li>
	<span class="prop">DAV:activity-checkout-set</span>
	activities are only used for working resources, and we do not
	record what working resources "exist".
      </li>
      <li>
	<span class="prop">DAV:activity-set</span> -
	activities are only used for working resources, so versions
	cannot be part of an activity.
	<div><i>maybe this should be defined as the empty set?</i></div>
      </li>
      <li>
	<span class="prop">DAV:version-controlled-binding-set</span> -
	we do not have version history resources to include in the
	property.
      </li>
    </ul>

    <h3>OPTIONS</h3>
    <p>
      The <span class="method">OPTIONS</span> request will signal that
      it supports the following DAV features:
    </p>
    <ul>
      <li><code>1</code></li>
      <li><code>2</code></li>
      <li><code>version-control</code></li>
      <li><code>checkout</code></li>
      <li><code>working-resource</code></li>
      <li><code>merge</code></li>
      <li><code>baseline</code></li>
      <li><code>activity</code></li>
      <li><code>version-controlled-collection</code></li>
    </ul>

    <h3>Reports</h3>
    <p>
      The <span class="prop">DAV:supported-report-set</span> property
      will signal support for the following reports:
    </p>
    <ul>
      <li>svn:update-report</li>
      <li>svn:log-report</li>
    </ul>
    <p>
      These reports are available only on the "public" resources (the
      VCRs). They are not available on the resources within the
      <span class="url">$svn/</span> area.
    </p>

    <h3>Notes, reminders</h3>
    <p>
      Discuss timeouts and auto-purge of activities (and the related
      working resources).
      <br/>
      Discuss the activity database maintained by mod_dav_svn.
      <br/>
      Discuss other implementation details of ra_dav and mod_dav_svn.
    </p>

    <h3><a name="rationale">Appendix A: Rationale</a></h3>
    <p>
      Several times, people have asked, "Why choose
      HTTP/WebDAV/DeltaV? That seems awfully bloated and
      ill-suited. Why didn't you design a custom, well-tuned protocol?
      Or maybe use the CVS protocol?" Listed below are a number of
      reasons for our choice of WebDAV as our network protocol.
    </p>

    <blockquote>
      <p><small>
	While this list could certainly be expanded with more reasons
	(and to be fair, with a list of reasons why WebDAV was a poor
	choice), it certainly demonstrates the basic reasons for our
	choice.
      </small></p>
    </blockquote>

    <blockquote>
      <p><small>
	Note: this list came from an email note, so the tone and point
	of view might be a bit off. Further word-smithing is
	welcome...
      </small></p>
    </blockquote>

    <h3>Builtin web browsing of the repository</h3>
    <blockquote>
      <p>For example, take a
      look at:
      <a href="http://svn.apache.org/repos/asf/subversion/trunk/README">http://svn.apache.org/repos/asf/subversion/trunk/README</a>
      (that's the HEAD right there; we also have URLs for every
      previous revision of every file)</p>
    </blockquote>

    <h3>DAV-based browsing</h3>
    <blockquote>
      <p>Use Web Folders or WebDrive or somesuch on
      your Windows box (or Windows XP's native DAV mounts) to browse
      the SVN repository with Windows Explorer. Mac OS X has builtin
      DAV server mounting. Nautilus has DAV capabilities. Then you
      have your Open Source tools such as cadaver, Goliath, etc.</p>
    </blockquote>

    <h3>People can use existing libraries</h3>
    <blockquote>
      <p>I couldn't even begin to count the number of HTTP tools and
      libraries available. If we had designed our own protocol, then
      we would have /none/ of those benefits. Heck, two HTTP library
      implementors (Joe Orton of Neon, and Daniel Stenberg of CURL)
      are regulars here. we wouldn't get that benefit. I've used
      Python's httplib (and a davlib of my own) to do a lot of testing
      of our server. No need to go and roll new protocol libraries.</p>
    </blockquote>

    <h3>Existing tools</h3>
    <blockquote>
      <p>One word: Ethereal :-) When we capture network traces, Ethereal
      already knows about HTTP. It's quite nice, but I know there are
      even better ones out there. But we also have other tools like
      squid and other (caching) proxies (see the next item).</p>
    </blockquote>

    <h3>Caching proxies</h3>
    <blockquote>
      <p>Subversion will work great with caching proxies. There is no
      longer a need for specialized tools like "cvsup". Just drop in a
      caching proxy, and you've already got your distributed read-only
      repository. That European dev team can just drop in the cache
      between them and the US server and their checkouts/updates will
      get cached for the benefit of the other team members. Commits
      will flow through, back to the US-based server.</p>
    </blockquote>

    <h3>Sophisticated and broad-choice authentication</h3>
    <blockquote>
      <p>We don't have to reimplement an authentication scheme for a new
      protocol. We can use all of the various schemes that have been
      defined for HTTP. Ever look at the CVS protocol? Ever see the "I
      Love You" or "I Hate You" lines? :-) That is all part of
      creating a new authentication scheme. But we get to use SSL and
      certificate-based auth if we want. Kerberos. NTLM. or even just
      simple Basic or Digest. And our users can come from text files,
      database, LDAP, or PAM. We don't have to reinvent the wheel cuz
      it is all available for Apache already.</p>
    </blockquote>

    <h3>Awesome network server</h3>
    <blockquote>
      <p>We don't have to worry about how to portably set TCP_CORK for
      optimal network packets. We don't have to worry about when
      sendfile() makes sense, or if it is available. We don't have to
      worry about dropped client connections, how to best use threads
      and processes to scale, request management, monitoring, logging,
      etc. Apache gives us all of that and a ton more. I *really*
      would not want to do that through xinetd. I mean... setting
      TCP_CORK on stdout? freaky :-)</p>
    </blockquote>

    <h3>Well-defined on-wire compression</h3>
    <blockquote>
      <p>We already have on-wire compression, similar to CVS's "-z#"
      switch. And we didn't do anything. The client library and server
      that we use just support it automatically for us, according to
      RFC 2616.</p>
    </blockquote>

    <h3>Future interoperability</h3>
    <blockquote>
      <p>In the future, we'll be able to interoperate with a multitude of
      IDEs and other WebDAV/DeltaV clients. As DeltaV becomes more
      prevalent, IDEs could very well use it for source code
      management, and we'll be right there without needing to write
      some MS/SCC library to interface to the tool.</p>
    </blockquote>

    <hr/>
    <address><a href="mailto:gstein@lyra.org">Greg Stein</a></address>
<!-- Created: Thu Aug 10 19:14:20 PDT 2000 -->
<!-- hhmts start -->
Last modified: Fri Jan 25 12:54:20 PST 2002
<!-- hhmts end -->

</div>
</body>
</html>